Control of Power Prosumer Based on Swarm Intelligence Algorithms

Page created by Max Chandler
 
CONTINUE READING
Control of Power Prosumer Based on Swarm Intelligence Algorithms
E3S Web of Conferences 209, 02020 (2020)                                                                https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

         Control of Power Prosumer Based on Swarm Intelligence
         Algorithms
         Pavel Matrenin1*, Vadim Manusov1, and Dmitry Antonenkov1
         1
          Department of Industrial Power Supply Systems, Novosibirsk State Technical University, Novosibirsk, Russia

                       Abstract. The development of renewable energy and Smart Grid leads to the emergence of prosumers or
                       power generating consumers, which are involved in the processes of bidirectional exchange of electricity
                       and information. The work is devoted to the problem of optimal control of a power generating consumer in
                       Smart Grid. The distinctive research features are the solution of the optimal control problem in conditions of
                       difficult prediction of wind power plant generation, the usage of Swarm Intelligence algorithms to build a
                       system of control rules, and the study of the obtained models on data from two different generating
                       consumers: one on about Russky Island, the second on Popov Island (Far East). We selected a list of priority
                       rules as a decision-making model and applied Particle Swarm Optimization, Bees Algorithm, and Firefly
                       Optimization to build and optimize this model. The computer modeling with the usage of two mounts
                       dataset showed that the proposed approach could significantly increase the revenue of the generating
                       consumers considered.

       1 Introduction                                                             A number of articles propose a stochastic game
                                                                              approach for the problem of energy trading between
       The development of renewable energy and Smart Grid                     smart grid prosumer. L. Ma et al. [6] used the energy
       leads to the emergence of prosumers or power generating                management model on cooperative game theory. S.R.
       consumers (GC), which are involved in the processes of                 Etesami et al. [9] formulated the interaction among
       bidirectional exchange of electricity and information [1,              prosumers as a stochastic game, in which each prosumer
       2]. GC needs to control not only electrical load but also              seeks to maximize its payoff, in terms of revenues and
       the flow of generated power. It significantly increases                proposed an optimal strategy for utility companies. The
       the complexity of its control tasks [3, 4].                            Stackelberg game approach for Smart Grid Energy
           The problem of the optimal GC control has a number                 Management (Energy sharing management) was
       of issues that lead to high complexity:                                considered at [10, 11].
           • GC operates under conditions of stochastic change                    Such management allows taking into account data on
       in the generation of electricity by renewable sources and,             all participants in the distributed electric power system,
       to a lesser extent, of its consumption;                                but there is a risk associated with the high level of
           • the control problem has a high dimensionality of                 centralization of the control system. Thus, modern
       the solution search space;                                             studies primarily consider the principles of constructing
           • the objective function is not an analytical                      the entire Smart Grid power system and the interaction
       expression, is need to be calculated algorithmically.                  rules for multiple GCs. Out research focuses on
           Much modern research has been devoted to optimal                   optimizing the control rules for a single GC with a
       control in Smart Grid networks with distributed                        difficult prediction of generation using Swarm
       generation and renewable energy sources. However, the                  Intelligence (SI) algorithms.
       optimal control is carried out at the level of a                           The SI algorithms are known to effectively solve
       supersystem in them, and not individual GC. The                        large-scale nonlinear optimization problems, including
       frameworks to real-time coordinate load scheduling,                    problems of power systems. The most commonly used
       sharing, trading were considered at studies [5, 6]. A.C.               SI algorithm is Particle Swarm Optimization (PSO);
       Luna et al. [7] proposed an energy management system                   paper [12] provides a comprehensive survey on the
       for coordinating the operation of distributed household                usage PSO for power system applications. Other SI
       prosumers with renewable energy sources. H. Mortaji et                 algorithms are also applied to different optimization
       al. [8] proposed smart-direct load control and load                    problems in power system design and control [13-15]. In
       shedding based on autoregressive integrated moving                     this research, three SI algorithms: PSO, Bees algorithm
       average time-series prediction model and Internet of                   (BA), and Firefly optimization (FFO).
       Things concept.

         *
             Corresponding author: pavel.matrenin@gmail.com

  © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0
  (http://creativecommons.org/licenses/by/4.0/).
Control of Power Prosumer Based on Swarm Intelligence Algorithms
E3S Web of Conferences 209, 02020 (2020)                                                         https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       2 The Problem Statement
       2.1 GC Power System
       In this research, we considered two large GC: the power
       system of Russky Island and the power system of Popov
       Island. Both islands are located in Peter the Great Gulf in
       the East Sea (Fig. 1). High wind speed makes it possible
       to create wind power plants up to 16 MW on Russky
       Island and up to 20 MW on Popov Island [16].
           Russky Island belongs to the territorial composition
       of Vladivostok. It is located about two kilometers from
       the coast in Peter the Great Gulf, which is part of the Sea
       of Japan (the smallest distance between the continental
       part and the island is 800 meters). Russky Island is
       separated from the Muravyov-Amursky Peninsula by the
       East Bosphorus. From the west, the island is washed by
       the waters of the Amur Gulf, and from the east and south
                                                                         Fig. 1. Russky and Popova Islands.
       side by the waters of the Ussuri Gulf. In the southwest,
       the island is separated from the other Popov island by the
       Stark Strait.
           The island is 97.6 km2, its length is about 18 km, and
       its width is about 13 km. The population of the island is
       approximately 25,000 inhabitants.
           Popova Island (named after Admiral A.A. Popova) is
       located in Peter the Great Gulf of the Sea of Japan, 20
       km from Vladivostok and 0.5 km southwest of Russky
       Island. About 3,000 people live on the island, mainly in
       the two villages of Stark and Popova.
           Fig. 2 and Fig 3. show curves of own consumption of           Fig. 2. Consumption and generation of Russky GC.
       Russky GC and Popova GC and possible wind power
       generation of GCs according to the estimates of [16]. Fig
       4. shows the results of adding these curves. The
       demonstrated fragment of data corresponds to twenty
       days from 01.06.2017.
           From the charts above it is possible to notice the
       following:
       • the forms of the electricity generation curves are very
       close since both islands are located very close, and their
       wind speeds are also close;
       • the load curves concerning to generation are radically
       different; the Russky GC always has a deficit, the                Fig. 3. Consumption and generation of Popova GC.
       Popova GC still has a surplus;
       • in case of consideration of two consumers together, we
       have the third type of load / generation pattern – in
       general, the GC system has a deficit, but sometimes
       there is a surplus.
           Thus, different management strategies may be
       required depending on the characteristics of GC or GC
       hub.

           2.2 Optimal Control

       The task of optimal control is to create a control system         Fig. 4. Aggregated consumption and generation of both GC.
       (subject) that implements a sequence of actions on a
       controlled object in the environment to achieve the best              The state of the controlled object is characterized by
       possible quality specified by one or more criteria (Fig.          a set of parameters that can change over time:
       5); the controlled object is a specific part of the world
       around which the control subject can purposefully                                S(t) = {s1(t), s2(t), ..., sn(t)}.           (1)
       influence [17]. Control always occurs during a certain
                                                                         Thus, there is a vector of functions. Each function shows
       period of time, while the controlled object passes from
                                                                         the parameter changing over time. These functions in the
       one state to another.

                                                                     2
Control of Power Prosumer Based on Swarm Intelligence Algorithms
E3S Web of Conferences 209, 02020 (2020)                                                                             https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       explicit form are unknown. In addition, there is a control                     cannot usually be obtained, especially integral of this
       system that provides control.                                                  function. But it is possible to calculate the function
                                                                                      algorithmically. In the case of GC control, this function
                                                                                      is piecewise continuous, since the time step is 1 hour.
                                                                                      The task (3) can be written without an integral, in the
                                                                                      form of a sum, and the function f(t, S(t), A(t)) is nothing
                                                                                      more than the difference between the revenues from the
                                                                                      sale of electricity of a GC and the costs of its purchase,
                                                                                      generation, and power storage in all hours into the time
                                                                                      period. However, even in this case, the analytical
       Fig. 5. Interaction between a subject and an object.                           expression for f(t, S(t), A(t)) is difficult to write, since the
                                                                                      price of electricity is a piecewise constant function, the
          The control can also be defined as a vector of                              exchange of electricity with a neighboring GC supply
       functions:                                                                     depends on its state and controlling them. Thus, the
                                                                                      calculation of the value of f(t, S(t), A(t)) should be
                            A(t) = {a1(t), a2(t), ..., am(t)}.              (2)       performed algorithmically:
                                                                                                                           T
          The notations S from “state” and A from “action” are                                Aopt (t )  arg min  revenue  t , S  t  , A  t     (4)
       used.                                                                                              A ( t ) A pos t  t0

          The optimal control problem, in general, can be
       written as follows:
                                                 tT

                      Aopt (t )  arg min  f  t , S  t  , A  t  dt   (3)
                                   A ( t ) A pos t
                                                   0

       • Aopt(t) is the required optimal control; it defines values
       of the control parameters at each time moment (in the
       considered task, when and how much GC must sell or
       buy, charge or discharge);
       • Apos is the area of permissible values of control
       parameters;
       • f(t, S(t), A(t)) is a continuous-time cost function, in the
       considered task, it gets GC’s total electricity costs:                         Fig. 6. Sample of daily state-action curves of Russky GC.
       purchases from the own generation + purchases from the
       other GC and an external power system – sale to other                          3 The Research Method
       GC and the external power system.
       • t0 and tT are the period of time considered.
                                                                                      3.1 Heuristic-Rule-based control
       2.3 GC Optimal Control task                                                    All possible control actions could be described by
                                                                                      dividing them into four groups. The following
       For GC, the state parameters can be defined as follows:                        designation is used:
       • own consumption, MWh (s1);                                                   • power_wind – GC wind power plant generation at the
       • wind power plants generation, MWh (s2);                                      considered hour;
       • charge level of power storage, MWh (s3).                                     • power_gc – GC consumption at the considered hour;
             Control parameters can be defined as follows:                            • dif – the difference between the GC generation and
       • the amount of electricity that is currently exchanged by                     consumption at the considered hour;
       the GC with an external power system (purchase or sale),                       • storage – the amount of energy that needs to be
       MWh (a1);                                                                      charged (> 0) or discharged (
E3S Web of Conferences 209, 02020 (2020)                                                          https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       • sale_buy – the amount of energy that is sold (> 0) or            has four balance factors: buy_unload, sale_unload,
       purchased (                  the procedure for their verification and compliance, that
       consumption).                                                      is, rule priorities. Decision making begins with checking
               1.1. dif = power_wind - power_gc;                          of the highest priority rule. If its condition is satisfied,
               1.2. storage = min(max_storage – now_storage,              then the corresponding action of this rule is
               max_storage_h, dif);                                       implemented. Otherwise, the next priority rule is
               1.3. storage = storage * sale_storage;                     checked, and so on until the end of the rule list. The
               1.4. now_storage = now_storage + storage;                  conditions are designed in such a way that when you go
               1.5. sale_buy = dif – storage.                             through the list of rules, you will surely find one whose
           2. Charge_Buy:                                                 condition will be satisfied. As a result, to build a
               2.1. dif = power_wind - power_gc;                          controller, it is necessary to determine the order of the
               2.2. storage = min(max_storage – now_storage,              rules by setting priorities (pri) and the tuned parameters
               max_storage_hour);                                         specified above:
               2.3. storage = storage * buy_storage;
               2.4. now_storage = now_storage + storage;                       Solution = [pr1, …, pr12, buy_unload, sale_unload,
               2.5 sale_buy = dif – storage.                                    buy_storage, sale_storage, time1, … , time4]
           3. Discharge_Sell:
               3.1. dif = power_wind - power_gc;                          3.2 Swarm Intelligence
               3.2. storage = now_storage;
               3.3. storage = storage * sale_unload;                      It is not always possible to determine the Swarm
               3.4. now_storage = now_storage – storage;                  Intelligence algorithm that is most suitable for a solved
               3.5. sale_buy = dif + storage.                             task [15]. Therefore, the use of only one algorithm can
           4. Discharge_Buy (it’s possible if generation <                give a solution whose effectiveness is not satisfactory for
       consumption):                                                      the optimization criterion. In this case, the researcher
               4.1. dif = power_wind - power_gc;                          cannot determine the effectiveness without using other
               4.2. storage = min(–dif, now_storage);                     algorithms for comparison. Therefore, three Swarm
               4.3. storage = storage * buy_unload;                       Intelligence algorithms were applied: the Particle Swarm
               4.4. now_storage = now_storage – storage;                  Optimization (PSO) algorithm, the Firefly Optimization
               4.5. sale_buy = storage – dif.                             (FFO) algorithm, and the Bees algorithm (BA) (not
           The choice of actions should depend on the state of            Artificial Bee Colony Optimization). Descriptions of the
       the GC, but it is enough to get answers to two questions.          algorithms precisely in the form in which they are
       The first is connected with determining whether the GC             applied in this research are given in.
       is in a state of excess or deficiency of energy? The
       second is also related to the fact that the price of
                                                                          3.2.1 Particle Swarm Optimization
       electricity changes throughout the day. Although various
       billing schemes are possible, a two-zone tariff is                 The Particle Swarm Optimization algorithm was first
       considered in this research, the daily tax is from 7 a.m. to       proposed by J. Kennedy and R. Eberhart in 1995 [18].
       11 p.m., and at other hours it is a night tax, cheaper one.        Then it was improved by Kennedy, Eberhart, and Shi
       Thus, it’s needed to get answers to the questions:                 [19]. PSO is based on a bird flocks’ behavior. A flock
           1) Excluding accumulation, does the generation of              acts coordinated according to a number of simple rules.
       the GC wind power plant more than the GC consumption               Every bird (called particle) coordinates own movements
       (diff> 0)?                                                         with the movements of whole flocks. In the PSO
           2) Is there a special time period now?                         algorithm, every particle is denoted by a position vector,
           The GC control takes into account the possibility of           a velocity vector, and a value of the criterion. The
       using two intervals as special periods (from time1 to              vectors of position and velocity of all particles are
       time2 and from time3 to time4), the values of the                  updated according to a number of rules taking into
       boundaries of the time intervals are parameters adjusted           account the best position of a particle, and the best
       during the optimization process.                                   position of the whole swarm. Also, the algorithm uses
           As a result, we have four possible cases at each hour:         inertia weights of the particles, velocities restriction and
       • (diff < 0) AND NOT (special_time_period);                        the stochastic deviations.
       • (diff > 0) AND NOT (special_time_period);                             According to the scheme of the swarm algorithms
       • (diff < 0) AND (special_time_period);                            description [15], the PSO algorithm may be represented
       • (diff > 0) AND (special_time_period).                            by a tuple {S, M, A, P, I, O}.
           When creating a GC control based on rules, we get                   1. A set of particles (particles) S = {s1, s2,…,s|S| }, |S|
       12 rules of the form IF , THEN                  is number of particles. At j-th iteration i-th particle is
           The number of rules is 12 since the second and third           characterized by the state sij = {Xij,Vij, Xbestij}, where Xij =
       actions can be performed under any of the four                     {x1ij, x2ij,…, xlij} is the variable parameter vector (particle
       conditions, and the first and fourth under two conditions          position), Vij = {v1ij, v2ij,…, vlij} is the velocity vector,
       (2 * 4 + 2 * 2 = 12). In addition, the GC control model

                                                                      4
E3S Web of Conferences 209, 02020 (2020)                                                             https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       Xbestij = {b1ij, b2ij,…, blij} are the best (by value) fitness-       the more bees go to it. However, the bees can randomly
       functions of the particle position among all the positions            deviate from the chosen direction. After the return of all
       it took during the algorithm operation from the 1 st to the           the bees in the hive, information exchange and sending
       j-th iterations, l is the number of variable parameters.              of bees again.
            2. Means of indirect exchange is vector M = Xbestj is                 According to the description scheme of swarm
       the best value of the variable parameters vector derived              algorithms, the BA algorithm may be represented by a
       among all particles from the 1st to the j-th iterations of            tuple {S, M, A, P, I, O}.
       the algorithm.                                                             1. A set of particles (bees) S = {s1, s2, …, s|S|}. At the
            3. Algorithm A describes the steps of the PSO                    j-th iteration the i-th particle is characterized by the state
       algorithm.                                                            sij = {Xij}, where Xij= {x1ij, x2ij, …, xlij,}is the variable
            3.1. Generation of initial population (iteration                 parameters vector (the particle position), l is the
       number j = 1):                                                        dimensionality of the solution search space.
            Xi1 ← random(0,1), i = 1, …, |S|,                                     2. Means of indirect exchange M is a list of the best
            Vi1 ← random(0, vmax), i = 1,…|S|,                               and perspective positions found in the j-th iteration, M =
            Xbesti1 ← Xij, i = 1, …|S|,                                      {Nijb, Nkjg}, i = 1, ..., nb, k = 1, ..., ng.
       where random(0, 1) is the vector of random numbers                         3. Algorithm A describes the steps of the BA
       with dimensionality l (dimensionality of solution search              algorithm.
       space) uniformly distributed from 0 to 1.                                  3.1. Generation of initial population (j=1) is fulfilled
            3.2. Calculation of fitness- functions. The criterion            only for a subset of particles termed scouts:
       calculation takes place in the mathematical model of the                   Xi1 ← random(0,1), i = 1, …, ns,
       problem where Xij vectors are entered from algorithms                      where ns is the number of scout particles. Other
       and the results are returned to the algorithm through the             particles are considered as inactive this time (only at the
       interface {I, O}.                                                     first iteration).
            Xbestij ← Xij | f(Xbestij)1 , i = 1, …, |S|.                              the lists of the best and perspective positions M = (Nijb,
            Xij+1 ← 0| Xij+1
E3S Web of Conferences 209, 02020 (2020)                                                              https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       the total number of the swarm particles (|S| = ns + nbсb +                   Where random ∈ [0, 1], and G(X) is used in this case
       ngсg). The selection of the algorithm parameters heavily                as the predicate showing if X belongs the area of
       affects the quality of derived solutions, so in order to                admissible solutions.
       increase the algorithm efficiency it is necessary to adapt                   The function v(Xij, Xkj) defines the attractiveness of k
       parameters.                                                             particle for i particle with j algorithm iteration:
                                                                                    v(Xij, Xkj) ← β·(1 + γ·r(Xij, Xkj))-1
                                                                                     where r(Xij, Xkj) is Cartesian distance between
       3.2.3 Firefly Optimization
                                                                               particles.
       Firefly Optimization was proposed by Xin-She Yang in                         3.4. If at the j-th iteration a stop-condition is
       2010 [21]. This algorithm as all Swarm Intelligence                     satisfied, then the value Xbesti is transmitted to output O,
       algorithms is based on the particles (fireflies) movement               or the transition to iteration 3.2 takes place.
       in the search searching space. Let’s consider the                            4. Vector P = {α, β, γ} are coefficients of the
       objective function minimum problem of the following                     algorithm. Coefficient α determines the influence degree
       type f(X), where X is a vector of varied parameters which               of stochastic algorithm nature. Coefficient β sets the
       can get the values from some D area. Each particle is                   degree of attraction between particles with zero distance
       specified by the value of X parameter and value of an                   between them i.e. defines the particle’s mutual influence.
       optimized function f(X). Thus, the particle is a feasible               Coefficient γ controls the dependence of attraction on the
       solution of the considered optimization problem.                        distance between particles.
            As the algorithm is based on watching for fly’s
       behavior, each particle is considered to see the “light”                3.3 Application the Swarm                     Intelligence
       from their neighbors, but the brightness of the “light”                 Algorithm for GC optimal control
       depends on the distance between particles. For the
       process of solution finding to be converged to the                      For applying SI algorithms, it is necessary to determine
       optimum, each particle in its movement takes into                       the mapping of the particle coordinate (X) in the search
       account only those neighbors having a better value of                   space solution to the solutions of the solved task. In this
       f(X) criterion. But for the algorithm not to degenerate                 case, the solution is the control actions A(t), as shown in
       into greedy heuristics, it is necessary to have particles’              expression (1). Each element of the vector X is bounded
       stochastic movement.                                                    from 0 to 1 [15]. The priorities are real numbers from 0.0
            According to the description scheme of swarm                       to 1.0, so pri = xi, i = 1, ..., 12. The parameters
       algorithms, the FFO algorithm may be represented by a                   buy_unload, sale_unload, buy_storage, sale_storage
       tuple {S, M, A, P, I, O}.                                               also take values from 0.0 to 1.0, so they are mapped in
            1. Set of particles (fire-flies). S = {s1, s2, …, s|S|}, |S|       the same way. To set values of time1, ..., time4, we use
       is a number of particles. At iteration j the ith particle is            rounded down values of 24x17, …, 24x20.
       specified by the state sij = {Xij}, where Xij = {x1ij, x2ij, …,             The FFO algorithm requires comparing each particle
       xlij} is a vector of the varied parameters (particle’s                  to each other, so the number of operations quadratically
       position), l is a number of the varied parameters.                      depends on the number of particles. The PSO and BA
            2. Means of indirect exchange is vector M is                       have a linear relationship. We reduce the number of FFO
       particles’ brightness.                                                  particles to equalize the calculation time. At the same
            M = {f(X1j), f(X2j), f(X|s|j)}                                     time, we increase the number of iterations of the FFO
            Brightness is determined by the optimality criterion.              algorithm to equalize the number of calculations of the
       This vector ensures the indirect experience exchange                    objective function. As a result, the number of particles is
       among particles.                                                        reduced four times, and the number of iterations is
            3. Algorithm A describes the steps of the ACO                      increased four times compared to the PSO algorithm and
       algorithm.                                                              the BA. The parameters of the SI algorithms are given in
            3.1. Generation of initial population (iteration                   Table 1.
       number j = 1):
            Xi1 ← random(G(X)), i = 1, …, |S|,                                           Table 1. Parameters of the SI algorithms
            where random(G(X)) is a vector of equally
       distributed random variables meeting the restrictions of                Alg.    Particles   Iteration      Heuristic coefficients
       searching space.
            3.2. Calculation of fitness-functions. The criterion                                                     α1 = 1.5 α2 = 1.5,
       calculation takes place in the mathematical model of the                PSO        200        500
                                                                                                                      ω = 0.7, β = 0.5
       problem where Xij vectors are entered from algorithms                                                      ns = 60, nb = 6, ng = 1,
       and the results are returned to the algorithm through the                BA        200        500             сb = 20, cg = 20,
       interface {I, O}.                                                                                           rad = 0.01, rx = 0.05
            mij ← f(Xij), i = 1, …, |S|
            Xjbest ← Xij | f(Xij) ≤ f(Xjbest)                                  FFO        50         2000         α = 0.05, β = 1, γ = 0.5
            3.3 Particles’ movement:
            Xij+1 ← Xij + v(Xij, Xkj) · (Xij – Xkj) + α·random(0, 1) |
       mkj ≤ mij , i, k = 1, …, |S|, i ≠ k,
            if G(Xij+1) = 0, Xij+1 ← Xij , i = 1, …, |S|,

                                                                           6
E3S Web of Conferences 209, 02020 (2020)                                                          https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       4 Results and Discussion                                           For the GC of Russky Island, there is a situation of
                                                                          electricity shortage, so control does not make a large
                                                                          contribution. For the Popova Island GC, on the contrary,
       4.1 Computational Experiment                                       there is an excess of electricity; therefore, it is necessary
       Computational experiments were carried out while                   to determine the best moments for the sale of electricity
       considering the GC of Russky and Popov Islands (GCR,               and the balance between sale and accumulation.
       GCP, respectively). Table 2 shows the prices used in the               The most interesting situation is when a joint system
       simulation.                                                        of two GCs is controlled together. First, in this case, the
                                                                          profile is more complicated (Fig. 4), since there are
                  Table 2. Prices used in the simulation                  moments of both excess and shortage of electricity.
                                                                          Secondly, the power storage capacity is two times higher
                                   Price,             Price,              due to the combination of storages of both GC in a single
              Power flow
                                rubel / MWh         $ / MWh               power system. Thirdly, the combined GC has a higher
            Wind generation          500              6,67                generation and consumption, since, it is evident that all
             Power storage                                                quantitative indicators will be more top.
                                     100              1,33                    The results of all applied SI algorithms are very
              discharging
            Sale (daily rate)        3200             42,7                close, even without adjusting the heuristic parameters. It
                                                                          can be explained by the relatively low complexity of the
            Sale (daily rate)        1400             18,67               task from the point of view of optimization theory since
            Buy (daily rate)         2700             36,00
                                                                          there are not many control options.

            Buy (daily rate)         900              12,00

           To evaluate the effectiveness of the rule-based
       control model built by the Swarm algorithms, we
       compared them with a base constructed manually by an
       expert. The main advantage of using SI is an automatic
       adaptation to the profiles of production and consumption
       of each GC. Therefore, the expert rules were constructed
       one time for the general case.
           Obviously, in the problem under consideration,
       control is impossible without power storage. Because
       without it, GC has to sell electricity at times of excess
       and buy at times of shortage, and there are no other
       options. Therefore, control efficiency is limited by the
       capacity of the power storage. Due to the limitation of
       capacity and the use of tariff simulation with only two            Fig. 7. Algorithms’ results. The histogram shows the monthly
       rates (day, night), the reduction in energy consumption            profit ($) of power storage usage with the different algorithms
       or the increase in income (this is the same thing) is not          of optimize control rule-based model. Left 4 bars shows results
       very large. Thus, for results clarification, we did not            for GCR, central 4 bars – for GCP, right 4 bars – for GCR+P.
       compare the absolute values of the cost of electricity
       according to criterion (4), but the benefits that control                           Table 3. Algorithms’ results
       gives regarding the situation without power storage.
           In section 2.1, three cases are considered: the control                                        Cost,    Monthly Profit,
                                                                               Case       Algorithm
       of a GC of Russky Island, GC of Popova Island, and an                                           thousands $  thousands $
       integrated system of both GCs. For each situation and                    GCR           No          580.3           -
       each algorithm, modeling was performed on data for two                   GCR         Expert        579.8        0.241
       summer months (Fig. 2-4).                                                GCR          PSO          579.4         0.43
                                                                                GCR          BA           579.5         0.43
                                                                                GCR          FFO          579.5         0.43
       4.2 Simulation results
                                                                                GCP           No          -41.6           -
       Table 3 shows the results. Each SI algorithm was                         GCP         Expert        -41.5          0
       launched 20 times, and in 17-18 launches out of 20 gave                  GCP          PSO          -42.8         0.6
       the same result. This result is used as a summary. The                   GCP          BA           -42,8         0.59
       results without the use of power storage are shown in                    GCP          FFO          -42,8         0.59
       rows with the label "No" in the "Algorithm" column, and                 GCR+P          No          452,0           -
       the results of control using the expert rules are labeled as            GCR+P        Expert        450,4         0.57
       "Expert". Also, Fig. 7 visualizes the benefits of using the             GCR+P         PSO          447,3         2.33
       rule-based model optimized by the SI.                                   GCR+P         BA           447,3         2.33
           It can be seen that the difference varies greatly                   GCR+P         FFO          447,3         2.33
       depending on the profile of production and consumption.

                                                                      7
E3S Web of Conferences 209, 02020 (2020)                                                       https://doi.org/10.1051/e3sconf/202020902020
ENERGY-21

       5 Conclusion                                                           Using Internet of Things in Smart Grid Demand
                                                                              Response Management. IEEE Transactions on
           In this research, we have applied Swarm Intelligence               Industry Applications, 53.6, pp. 5155-5163 (2017)
       algorithms to optimize the heuristic-rule-based control          9.    S.R. Etesami, W. Saad, N.B. Mandayam, H.V. Poor.
       model (optimize priorities of rules and numerical values               Stochastic Games for the Smart Grid Energy
       of the model coefficients) for optimal control of                      Management with Prospect Prosumers. IEEE
       generating consumers with wind power plants. The                       Transactions on Automatic Control, 63.8, pp. 2327-
       computer simulation showed that SI allowed to increase                 2342 (2018)
       the profit of power storage usage 1.8–4.1 times
       compared with the rules build by experts. The simulation         10.   G. El Rahi, et. al. Managing Price Uncertainty in
       results confirmed that it is appropriate to apply the SI               Prosumer-Centric Energy Trading: A Prospect-
       algorithms to increase accuracy of heuristic-rule-based                Theoretic       Stackelberg      Game     Approach.
       model, and perform adaptation to a given generating                    Transactions on Smart Grid, 10.1, pp. 702-713
       consumer.                                                              (2019)
          The proposed control model allows to get the robust           11.   N. Liu, X. Yu, C. Wang, J. Wang. Energy Sharing
       control in a particular situation, which can be easily                 Management for Microgrids with PV Prosumers: A
       transferred to other climatic conditions and GC features.              Stackelberg Game Approach. IEEE Transactions on
       For future work, we plan: firstly, to complicate a GC                  Industrial Informatics, 13.3, pp. 1088-1098 (2017)
       model and a model of GCs interaction; secondly, to               12.   Y. del Valle, et al. Particle Swarm Optimization:
       apply the Q-learning method for the optimal control of                 Basic Concepts, Variants and Applications in Power
       GC; thirdly, make comparisons with existing methods                    Systems. IEEE Transactions on Evolutionary
       using larger dataset.                                                  Computation, 12.2, pp. 171-195 (2008)
                                                                        13.   V.Z. Manusov, P.V. Matrenin, L.S. Atabaeva.
       This work is supported the Novosibirsk State Technical                 Firefly Algorithm to Optimal Distribution of
       University Development Program through the Project C20-20.             Reactive Power Compensation Units. International
                                                                              Journal of Electrical and Computer Engineering,
                                                                              8.3, pp. 1758-1765 (2018)
       References
                                                                        14.   V.Z. Manusov, et al Optimization of Power
       1.   C.W. Gellings. The Smart Grid: enabling energy                    Distribution Networks in Megacities. IOP
            efficiency and demand response. (Lilburn, CA:                     Conference Series: Earth and Environmental
            Fairmont Press, 2009).                                            Science. International Conference on Sustainable
       2.   V.Z. Manusov, N. Khasanzoda, P.V. Matrenin.                       Cities, 72, id 012019 (2017)
            Application of artificial intelligence methods to           15.   V.Z. Manusov, P.V. Matrenin, S.E. Kokin. Swarm
            energy management in Smart Grids (Novosibirsk,                    Intelligence Algorithms for The Problem of The
            2019) [In Russian]                                                Optimal Placement and Operation Control of
       3.   X. Fang, S. Misra, G. Xue, D. Yang. Managing                      Reactive Power Sources into Power Grids.
            smart grid information in the cloud: Opportunities                International Journal of Design & Nature and
            model and applications. IEEE Netw., 26.4, 32-38                   Ecodynamics, 12.1, pp.101-112 (2017)
            (2012).                                                     16.   N. Khasanzoda. Optimization of power consumption
       4.   R. Zafar, et. al. Prosumer based energy management                modes in intelligent networks with a two-way flow of
            and sharing in smart grid. Renewable and                          energy using artificial intelligence methods
            Sustainable Energy Reviews, 82.1, pp. 1675-1684                   (Novosibisk, 2019) [In Russian]
            (2018)                                                      17.   L.A. Rastrigin, Modern principles of management of
       5.   A.G. Azar, et. al. A Non-Cooperative Framework                    complex objects (Moskov, 1980) [In Russian]
            for Coordinating a Neighborhood of Distributed              18.   J. Kennedy, R. Eberhart. Particle swarm
            Prosumers. IEEE Transactions on Industrial                        optimization. IEEE International Conference on
            Informatics, 15.5, pp. 2523-2534 (2019)                           Neural Networks, Perth, WA, Australia. pp. 1942–
       6.   L. Ma, N. Liu, J. Zhang, L. Wang. Real-Time                       1948 (1995)
            Rolling Horizon Energy Management for the                   19.   R.C. Eberhart, Y. Shi. Particle swarm optimization:
            Energy-Hub-Coordinated Prosumer Community                         developments, applications and resources. Congress
            From a Cooperative Perspective. IEEE Transactions                 on Evolutionary Computation; Seoul, South Korea.
            on Power Systems, 34.2, pp. 1227-1242 (2019)                      pp. 81-86 (2001)
       7.   A.C. Luna, N.L. Diaz, M. Graells, J.C. Vasquez,             20.   D.T. Pham, et. al. The bees algorithm – a novel tool
            J.M. Guerrero. Cooperative energy management for                  for complex optimisation problems (Cardiff, UK,
            a cluster of households prosumers. IEEE                           2005)
            Transactions on Consumer Electronics, 62.3, pp.             21.   X. Yang. Firefly algorithm, Stochastic Test Function
            235-242 (2016)                                                    and Design Optimization. International Journal of
       8.   H. Mortaji, S. Siew, M. Moghavvemi, H. Almurib.                   Bio-Inspired Computation. 2.2, pp. 78-84 (2010)
            Load Shedding and Smart-Direct Load Control

                                                                    8
You can also read